Service
Power BI and data analytics
Power BI often gets used like Excel with nicer charts. I build dashboards the other way around: data model first, then measures, then visuals. The result refreshes itself and does not break at the next new data source.
The problem
What I commonly find: a Power BI report built on one giant flat table, DAX measures that are really calculated columns in disguise, and a refresh nobody wants to touch since the original author left. Every new metric means a new column in the raw table instead of a new measure in the model.
Data sources are often a patchwork: a SQL export here, a manually maintained spreadsheet there, a SharePoint folder with monthly CSV files. Without a clean star schema, every new source adds fragility, and nobody can explain why two metrics show slightly different numbers on two report pages.
Microsoft Fabric and Azure connectivity do not automatically fix this. Without a well-thought-out data model, the Excel chaos just moves into a more expensive cloud environment.
The solution
I build the data model first: a star schema with clearly separated fact and dimension tables, relationships instead of nested lookup logic, measures instead of calculated columns. That is the difference between a dashboard still understandable in a year and one nobody wants to extend after three months.
Power Query handles the connection to sources (SQL Server, Azure, SharePoint, spreadsheet exports), including query folding wherever the source supports it, and Microsoft Fabric/Lakehouse components where data volume or governance requirements justify it. DAX measures are documented and organized by topic, not dumped in an unsorted list.
Deliverables: a documented data model, reusable measures, a refresh plan (scheduled in the Power BI Service or via an on-premises gateway), and a README that explains where every metric comes from and why.
Typical use cases
- Replacing a manually maintained reporting spreadsheet with a Power BI dashboard that refreshes automatically
- Building a star schema for an existing, grown Power BI model with conflicting metrics
- Connecting SQL Server, Azure data sources, or Microsoft Fabric/Lakehouse to a Power BI model
- DAX audit: reviewing existing measures, fixing performance issues and incorrect context transitions
- Executive dashboards with clear KPI definitions and drill-down instead of an unconnected pile of charts
- Migrating Excel pivot reporting to Power BI, including a short team training session
Concrete benefits
- One data model, no copies: metrics agree across every report page
- Automatic refresh instead of manual spreadsheet upkeep, scheduled in the Power BI Service or gateway
- DAX measures are documented and understandable for a second analyst within one to two days
- Scales toward Microsoft Fabric once data volume or governance needs grow into it
- Fewer wrong numbers, because the logic lives once in the model instead of scattered across spreadsheet formulas
How we work together
Audit of the existing report or requirement (1-2 days)
Which metrics exist, where does the data come from, where do numbers conflict? Output is a prioritized plan for the data model and measures.
Build the data model and Power Query connections
Star schema, relationships, query-folding-friendly Power Query steps. Connects to Microsoft Fabric/Lakehouse or Azure sources where needed.
DAX measures and visuals
Measures documented by topic, report pages with clear navigation, drill-down where it helps, no pile of disconnected charts.
Handover with a refresh plan and walkthrough
README with metric provenance, scheduled refresh, and maintenance notes. 60 to 90 minute live walkthrough, recorded.
Frequently asked questions
Power BI vs. Excel: when is switching worth it?
As soon as several people interpret the same report differently, as soon as the report gets refreshed by hand every month, or as soon as several data sources come together. For a one-off analysis, Excel plus Power Query is often still enough without Power BI on top.
What does a typical Power BI project cost?
Billed at €100 per hour, with total effort depending heavily on the number of data sources and metrics. A pure DAX/model audit of an existing file is considerably cheaper than a dashboard rebuild. You get a real range after the discovery workshop.
Do we need Power BI Pro or is the free version enough?
Power BI Desktop for building is free. You need Pro once you want to share reports with others in the Power BI Service. For a single user without sharing, the free tier is often enough.
How does this relate to Microsoft Fabric?
Fabric pays off once data volume or governance requirements outgrow a classic Power BI model, for example a Lakehouse feeding several downstream consumers. For most mid-size dashboards, Power BI with a clean Power Query foundation is enough, Fabric is not an automatic upgrade.
Can we extend the model ourselves later?
Yes, if measures are documented and organized by topic. Someone with basic DAX knowledge can extend it within one to two days. Deeper questions (context transitions, performance on large models) are a good fit for a retainer.
/ Next step
Audit or rebuild your Power BI model?
Briefly describe which metrics are still assembled by hand today and where the data comes from. You get an honest read on where the data model should start.